Replying again since you updated in response to my questions =)
1. It's far faster. Installations and all operations take less time.
That hasn't been my experience, but have you checked the latest release? It switched the local installation information to sqlite, which seems to make things mighty fast.
2. It's far lighter. Sure disk space is cheap, but why waste if I can avoid it?
Because the "we won't replicate dependencies" thing won't survive another full OS X release cycle. Snow Leopard was a rushed release cycle that brought a quick refresh to the OS-supplied components, but it was a once-off and it's not going to happen again soon, especially given the focus on iOS.
3. ... And Github is Homebrew's secret weapon here. There are literally hundreds of us with forks, providing constant updates. Packages are updated far faster than Fink or MacPorts. Momentum counts.
That doesn't seem to match reality. Take the current case; MacPorts:
* Has been tracking Python2.7 releases since its alpha release 5 months ago.
* Provides ports for python 2.4, 2.5, 2.6, 2.7 and 3.1 (all of which are not always backwards compatible)
Homebrew is tracking only Python2.6, didn't have any Python2.7 submissions until yesterday, and hasn't integrated the change into mainline.
When they -do- integrate the change into mainline, everybody is going to have to deal with the headaches that invariably occur across minor releases of Python[1], rather than tracking the multiple versions they require, because homebrew has a naively simple and ultimately flawed policy of "only one version".
Most of the complexity in MacPorts and Fink is a direct result of understanding the complexity of the problem space over a decade. Homebrew will become just as complex as time reveals the necessity of many of these compromises.
1. It's far faster. Installations and all operations take less time.
That hasn't been my experience, but have you checked the latest release? It switched the local installation information to sqlite, which seems to make things mighty fast.
2. It's far lighter. Sure disk space is cheap, but why waste if I can avoid it?
Because the "we won't replicate dependencies" thing won't survive another full OS X release cycle. Snow Leopard was a rushed release cycle that brought a quick refresh to the OS-supplied components, but it was a once-off and it's not going to happen again soon, especially given the focus on iOS.
3. ... And Github is Homebrew's secret weapon here. There are literally hundreds of us with forks, providing constant updates. Packages are updated far faster than Fink or MacPorts. Momentum counts.
That doesn't seem to match reality. Take the current case; MacPorts:
* Has been tracking Python2.7 releases since its alpha release 5 months ago.
* Has years worths of institutional knowledge regarding patching Python to actually fully function (see the portfile here: http://trac.macports.org/browser/trunk/dports/lang/python27/...).
* Provides ports for python 2.4, 2.5, 2.6, 2.7 and 3.1 (all of which are not always backwards compatible)
Homebrew is tracking only Python2.6, didn't have any Python2.7 submissions until yesterday, and hasn't integrated the change into mainline.
When they -do- integrate the change into mainline, everybody is going to have to deal with the headaches that invariably occur across minor releases of Python[1], rather than tracking the multiple versions they require, because homebrew has a naively simple and ultimately flawed policy of "only one version".
Most of the complexity in MacPorts and Fink is a direct result of understanding the complexity of the problem space over a decade. Homebrew will become just as complex as time reveals the necessity of many of these compromises.
[1] As an example, http://twistedmatrix.com/trac/query?status=closed&group=...